I think that people will be surprised at how simple this machine really is. Even though I'm having problems building it, the basic idea is so simple that, as Karl commented, I cannot understand why it hasn't been discovered before. Bessler gives an amazing assortment of clues which all make sense once you have the whole picture but individually they seem to counter each other.
I still maintain my belief that five mechansisms are required and I'll try and explain why without giving too much away. If you assume that a piece of the mechanism has to fall at some point, then it is logical to think that the maximum benefit from that fall will be obtained from a right angled fall, i.e. 90 degrees. You could increase this up to 180 degrees, but half of that fall would be counter productive because .... think, which might be more effective? To start from twelve o'clock and fall to three, or start at three and fall to six o'clock. Any angle outside those two and you stray into the other angle's area of effectiveness.
What is the next whole number after four that would fulfill the need to have a continuous input from falling weights? Five.
If that is a confusing explanation, I apologise. I thought I'd have a go at explaining my conviction that Bessler used five mechanisms without giving away the solution. Maybe I have?